APH 2000, 58, 307-319:

Linking a breast cancer screening data base with a cancer registry in Antwerp, Belgium.

G. Van Hal, A. Thibaut, M. Matthyssen and J. Weyler

Keywords: Belgium, feasibility studies, mammary neoplasms, mass screening, medical record linkage, program evaluation, registries

The effect of well organized breast cancer screening on breast cancer mortality has been demonstrated for more than 20 years.  However, since the success of a screening programme is not unconditional, evaluation of its effect is always necessary.  For an effect evaluation of a breast cancer screening programme, a cancer registry is indispensable.  In several European countries evaluation of the effect of breast cancer screening can be performed usign routinely available cancer registry data.  The main objective of our study was to assess the possibilities and difficulties of linking a cancer registry with a screening data base in the Rupelstreek, Flanders/Belgium.

The study population comprised 50 to 70 year old women (n = 9,321), invited to the 1st round (1991-1992) and/or the 2nd round (1994) of the screening programme.  Of these women, 4,267 participated in at least one round (45,8%).  There were 41 women with a screening detected breast cancer and 107 women who sent back a reply card indicating they were breast cancer patients.

To collect as much information as possible on breast cancer patients, all GPs in the region (n = 60) were sent a questionnaire in 1996.  In addition, the Screening Data Base (SDB) was linked to the Antwerp Cancer Registry (ACR), which, especially in its early years, stressed the registration of breast cancer.

Six women with breast cancer could not be retrieved by the ACR of by the SDB and could only be traced thanks to the GP survey.

The linking of the ACR with the SDB showed that of the 148 breast cancer patients who were registered in the SDB, 75 were also recorded in the ACR (50,7%).  Furthermore, we could register 13 screening participants whose breast cancers only showed up in the meantime between 2 consecutive screening rounds and another 34 breast cancer patients only thanks to the ACR.

The linking of the SDB with the ACR seems very promising but is very labour intensive and until further notice, the privacy laws make it impossible to use nominative data for linking.  In addition, there is currently no watertight cancer registry.

Our study clearly indicated that the amount of manpower, time and financial support needed for permanent surveillance and quality control of organised breast cancer screening must certainly not be underestimated.